Re: Systemd files on a Raspberry Pi
On Mon, Feb 11, 2019 at 03:55:04AM +1100, Andrew McGlashan wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
>
> Hi,
>
> On 10/2/19 11:17 pm, Reco wrote:
> >> Okay, I've watched it now. I am not convinced that his idea of
> >> "create this or that yourself" is a fair retort.
> >
> > There are historical precedents. AIX's init inspired Solaris' SMF
> > which in turn inspired systemd.
>
> Not SMF -> systemd ... according to the video. More inspired by
> Apple's launchd
Which, in turn, has xml, central database, socket activation and very
rudimentary dependency resolution. I don't remember off-hand which one
came first, launchd or SMF.
> >> I still consider that systemd has caused considerable harm to
> >> Debian as well as the general Linux community.
> >
> > Those who do not want systemd are free not to use it. And Debian is
> > one of the few Linux distributions which provides such choice.
>
> If all goes to Pot, then choice will be lessened as projects make it
> too difficult to maintain versions of their code sans systemd.
That's up to his IBM overlords to decide. As usual for the typical
corporate development. So far they said nothing.
> >> The "Unix is dead" comment was interesting.
> >
> > AIX and Solaris (well, AIX mostly) are begging to differ. Free
> > software Unix (if such thing ever existed) - that's dead for sure.
>
> There is at least the BSD variants, are they really dead too?
I specifically mention "Unix". FreeBSD, which is the parent of all
modern BSDs, is free from AT&T code since '94, therefore is not a Unix.
BSDs may be twitching, but they ain't Unix.
> >> Has it really has become Linux or nothing?
> >
> > No. See SCO vs IBM case. Linux is not Unix, and never has been.
>
> We also have Luminos varieties... SmartOS and others.
They are as stillborn as their parent, OpenSolaris.
Also, see Netapp vs Nexenta case. That's what happens to OpenSolaris
offsprings once they leave their crib.
> >> It is a pity that Oracle has their licensing problems in relation
> >> to ZFS
> >
> > Please blame Sun Microsystems for *that*. Oracle's merely keeping
> > the status-quo.
>
> Yes, but Oracle could fix that if they wanted to.
Why would they? For Oracle any Operating System is a big launcher of
their database and assorted Java shovelware.
> The JAVA fiasco that Oracle hope to profit from royally from Google is
> another box of pox.
So called 'JAVA fiasco' had its share of hype, but that's it.
If you're looking for real fiascoes (sp?), search for HP vs Oracle case.
> >> and there are great alternative implementations now;
> >
> > You mean, ZFS-on-Linux? It's a fork of OpenSolaris' ZFS
> > implementation, not something that's written from scratch.
>
> Maybe so, but it isn't limited to Sun (or now Oracle systems).
> And it isn't limited to Linux as opposed to BSD variants or Luminos
> variants.
And here lies the irony. FreeBSD, OpenIndiana, Illumos - they all now
consider ZFS-on-Linux as upstream.
And, as Solaris 11.4 shows us, Oracle too can consider borrowing a
feature or two from ZFS-on-Linux.
The main problem is - ZFS is not native for Linux. It was (as still -
see Solaris porting layer kernel module - spl) written with Solaris
kernel facilities at mind.
Reco
Reply to: